Agentic AI era for the commerce landscape: Goldman Sachs sees agentic AI shifting commerce value toward platforms that capture consumer intent and enable trusted transactions

Overview

Goldman Sachs examines how agentic AI could alter product discovery, purchasing and distribution economics across the commerce landscape. Its central conclusion is that value should increasingly accrue to companies that capture commercial intent, retain consumer trust, secure merchant participation and provide the infrastructure needed to execute transactions safely.

Core views

Goldman Sachs frames agentic commerce as a long-duration, 3–5+ year transition rather than an immediate replacement for existing shopping behavior. Adoption is currently concentrated in discovery, research and recommendations, with measured progress expected toward end-to-end agentic shopping. The report estimates that $2.6 trillion of U.S. spending falls into high-likelihood categories—low-risk, recurring purchases—and that every 2% of agentic penetration of card-present spending could add roughly 1% to e-commerce growth. Adoption should arrive sooner where transaction failure has low consequences, while high-cost, highly personalized or return-prone purchases should take longer. The report’s overarching view is that value should follow consumer intent. AI interfaces may displace some traditional search and retailer-site discovery, but Goldman Sachs does not expect advertising to disappear. Instead, it expects monetization to migrate toward AI-native sponsored recommendations and other surfaces where commercial intent forms. This is compared with the desktop-to-mobile transition: near term, AI improves creative efficiency, campaign optimization and return on ad spend; over time, a larger portion of shopping intent may originate within AI interfaces. The key debate is whether this transition democratizes discovery for smaller merchants or instead produces new gatekeepers that charge for access to consumer intent. For retailers and marketplaces, the report sees a tension between agent-enabled discovery and preserving direct customer relationships. Large retailers may resist ceding first-party data, loyalty, cross-selling and retail-media economics to horizontal agents. At the same time, agents may improve discovery for smaller merchants, including Shopify merchants. Scaled retailers such as Amazon and Walmart may retain advantages because agents can optimize for price, product availability, delivery speed and transaction reliability. Walmart’s Sparky, launched in June 2025, is cited as an example: weekly active customers reportedly doubled year on year, grew 60% quarter on quarter, and generated a 40% higher average order value. The report argues that Walmart’s assortment, price, speed, fulfillment network and partnerships with AI platforms could support customer acquisition and incremental sales, while its data-sharing boundaries preserve longer-term customer data. Goldman Sachs expects a bifurcation among consumer categories and brands. For routine replenishment categories such as paper goods, cleaners, center-store food and apparel staples, agents can manage the full purchase cycle, substitute products when items are unavailable and prioritize value and availability. That can reduce brand loyalty and increase private-label pressure. By contrast, higher-involvement, recommendation-led categories—such as prestige beauty, skincare and innovation-led consumer products—retain a role for consumer judgment, efficacy claims, loyalty and direct-to-consumer engagement. The report argues that brands need machine-readable product attributes, differentiated features, storytelling, top-of-mind awareness and AI-engine optimization rather than relying on information buried in PDFs. It highlights SharkNinja because it operates in more considered, innovation-led categories with lower substitution risk, and Tapestry because brand equity and customer engagement may support preference where agents can readily surface substitutes. Payments are viewed constructively because agentic payments build on card infrastructure rather than bypassing it. Protocols from Google, Stripe, Visa and Mastercard are connecting agents, merchants and authorization processes, while liability standards remain unresolved. Goldman Sachs believes Visa and Mastercard are well positioned as networks scale their authorization, tokenization and value-added-service capabilities. Faster e-commerce growth, greater value-added-services attachment and fragmentation of spending across smaller merchants could support the networks. The report estimates that a 5% acceleration in Visa/Mastercard cybersecurity-related value-added services would accelerate value-added-services growth by 90 basis points and 200 basis points, respectively, and total net revenue growth by 20 basis points and 80 basis points, respectively. Modern consumer-finance and buy-now-pay-later issuers may also benefit if consumers favor more flexible, cheaper and simpler financing options. Trust infrastructure is a central requirement for adoption. Legitimate agents can make conventional fraud signals less reliable because activity may originate through shared cloud infrastructure, follow programmatic navigation patterns and occur at machine speed. Goldman Sachs therefore expects greater demand for verified agent identity, consumer authorization, merchant monitoring, fraud detection and governed real-time decisioning. It identifies four decision areas: validating consumers, assessing merchant legitimacy, detecting anomalous transactions and determining financing eligibility. Equifax is positioned through broad identity, business, taxpayer, payment-account, credit, income and employment data; TransUnion through persistent identity resolution and network intelligence; and FICO through borrower-risk scores, credit optimization, fraud detection and enterprise decisioning. Cloudflare is positioned at the merchant edge to distinguish authorized shopping agents from malicious bots, while Okta is positioned in identity, authorization and control of the data and actions available to agents. The same trust gap is also an adoption risk. The report states that widespread use requires clear rules on intent verification, erroneous purchases, fraud losses, chargebacks and returns. Payments fraud is specifically identified as a risk, with dispute abuse and card testing cited more often than refund abuse, instant-payments fraud and phishing attacks. The difficulty of distinguishing legitimate automation from malicious activity could increase cybersecurity spending, particularly for AI-enabled and pre-authorization detection. For live-event ticketing, Goldman Sachs expects different outcomes for primary and secondary platforms. Agents could reduce discovery friction by matching event seekers with relevant events by price, location, artist and venue, supporting higher sell-through and utilization for events with excess capacity. Primary ticketing platforms are viewed as better placed because they control differentiated inventory and rely less on customer acquisition. Secondary marketplaces could initially benefit if agentic channels lower acquisition costs below paid-search costs, but they face longer-term risks as AI platforms gain pricing power and charge referral, placement or commission fees. All-in pricing and AI-driven comparison can shift competition toward price and fulfillment reliability, pressuring service fees where comparable seats are listed across multiple marketplaces. Agent-led purchasing may also reduce ancillary attachment and the value of paid placement.

Analysis framework

Goldman Sachs begins with the adoption path for agentic shopping, then traces how a shift in commercial intent could affect e-commerce, advertising, retailers and consumer brands. It assesses payment and trust infrastructure by mapping the authorization, identity, fraud, merchant-verification and financing decisions required for agent-led transactions, and then applies the framework to selected companies and to primary versus secondary ticketing models.

Methodology notes

• Industry AnalysisUpstream-Midstream-Downstream Transmission

  Commerce-value-chain transmission

  The report follows the effect of agentic shopping from discovery and advertising through retailer and merchant economics, payment authorization, identity verification, fraud prevention and transaction execution.

• Industry AnalysisSubstitution-Effect Analysis

  Category-based substitution and consumer-involvement analysis

  The report distinguishes low-involvement, commoditized purchases—where agents can substitute products more readily—from considered, differentiated purchases where brand preference and consumer choice remain more important.

• Competition & strategyEconomic Moat and Competitive Advantage

  Competitive-position assessment

  Goldman Sachs evaluates beneficiaries based on distribution, consumer trust, merchant participation, transaction infrastructure and the ability to capture commercial intent.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• SharkNinja (SN)

  Highlighted as positively skewed to agentic discovery.

  Strengths

  Participates in considered, innovation-led categories with lower substitution risk as AI narrows recommendation sets.

  Comparison

  Positioned ahead of the average brand through a technology-forward approach.

• Tapestry (TPR)

  Highlighted as positively skewed to agentic discovery.

  Strengths

  Brand equity and customer engagement may sustain preference where agents can surface substitutes.

  Comparison

  Also described as technology-forward relative to the average brand.

  Risks

  Easy comparison among substitutes remains a challenge.

• Visa (V) and Mastercard (MA)

  Potential beneficiaries through card infrastructure and value-added services.

  Strengths

  Scale, existing authorization infrastructure, tokenization and potential cybersecurity-service attachment.

  Weaknesses

  Liability allocation for agentic transactions remains unresolved.

  Comparison

  Viewed as better equipped than new entrants to provide infrastructure during changing consumer behavior.

  Risks

  Fraud liability and technical requirements remain uncertain.

• Equifax (EFX), TransUnion (TRU) and Fair Isaac (FICO)

  Potential beneficiaries from agentic identity, fraud and credit-decisioning workflows.

  Strengths

  EFX offers broad verification data; TRU provides persistent identity and network intelligence; FICO operates in borrower-risk assessment and decisioning.

  Comparison

  The companies address different layers of identity, fraud and credit decisions.

• Cloudflare (NET) and Okta (OKTA)

  Potential beneficiaries from heightened agent identity, authorization and cybersecurity requirements.

  Strengths

  NET can authenticate and manage automated traffic at the merchant edge; OKTA can link agents to authenticated consumers and govern access.

  Comparison

  NET is focused on the merchant edge, while OKTA operates at the identity and authorization layer.

  Risks

  The evolving protocol and security environment may affect implementation.

• Live Nation Entertainment (LYV) and StubHub Holdings (STUB)

  Examples of primary and secondary ticketing exposure, respectively.

  Strengths

  Primary ticketing benefits from differentiated inventory and lower customer-acquisition reliance.

  Weaknesses

  Secondary inventory is more commoditized and exposed to price comparison.

  Comparison

  Goldman Sachs views primary ticketing platforms as better positioned than secondary marketplaces.

  Risks

  New distribution tolls, service-fee pressure, ancillary leakage and lower paid-placement value.

Key data

• High-likelihood U.S. spending categories$2.6tnEstimated U.S. spend in low-risk, recurring-purchase categories considered likely to adopt agentic commerce.
• E-commerce growth sensitivity~1%Estimated additional e-commerce growth for every 2% of agentic penetration of card-present spend.
• Transition duration3–5+ yearsGoldman Sachs characterizes agentic commerce as a long-duration transition.
• Walmart Sparky weekly active customersDoubled y/y; +60% q/qReported rapid adoption following its June 2025 launch.
• Walmart Sparky average order value+40%Reported AOV uplift associated with Sparky.
• Visa/Mastercard cybersecurity VAS sensitivity5% acceleration in cybersecurity-related VASEstimated to add 90bps/200bps to Visa/Mastercard VAS growth and 20bps/80bps to total net revenue growth, respectively.

Impact & implications

The report expects the economic value of shopping discovery to migrate toward businesses that own or monetize commercial intent and can deliver trusted, reliable execution. It is constructive on commerce infrastructure, card networks, modern credit issuers, identity and fraud decisioning providers, and cybersecurity vendors, while warning that commoditized retail categories and secondary ticketing marketplaces may face greater pricing and distribution pressure.

Risks

• Widespread adoption depends on clear rules for intent verification, erroneous purchases, fraud losses, chargebacks and returns.
• Traditional fraud signals may weaken when authorized agents operate through shared cloud infrastructure and at machine speed.
• Large retailers could lose first-party data, loyalty, cross-selling and retail-media economics if horizontal agents control discovery.
• Commoditized retail categories may face substitution pressure as agents prioritize price, availability and delivery speed.
• Secondary ticketing platforms may face fee pressure, new referral costs and weaker ancillary revenue as AI simplifies price comparison.

What to watch

• The pace at which agentic shopping expands beyond discovery, research and recommendations into full transaction execution.
• Emerging standards for agent authentication, consumer authorization, liability, chargebacks and returns.
• Whether AI platforms monetize control of commercial intent through sponsored recommendations, referral fees or other tolls.
• Walmart Sparky adoption, order-value trends and the role of AI partnerships in customer acquisition.
• Adoption of Google, Stripe, Visa and Mastercard protocols for agent-to-merchant payments.
• Demand for identity verification, fraud decisioning, bot management and cybersecurity as automated commerce grows.